Barton 2600+ Unstable

I've had this chip in for a couple of days now.

Running at stock speed but have instability...I.E. random reboots, even without applying load sometimes, but always under load. Idle temp is 40C, load 47-48C (I'll get better cooling after I get back into town after Thanksgiving), but right now I'm running stock heatsink and fan. System temp is a steady 17C.

I tried various combos of voltage increases...cpu, ram, chipset, but no help. So now am running system defaults.

I noticed in bios and MBM that cpu seems to be undervolted....default is 1.65, but reads 1.63-1.64 in bios and MBM...it fluctuates, is this normal?

Even fter raising core voltage to 1.675 I get readings of 1.63-1.65, again fluctuating. Again, just curious if this is normal or is it a problem with my crappy power supply?

I have an Arrow 500W psu. The 3.3v line reads 3.57v, 5.00v reads 4.97v, 12v fluctuates between 11.55v and 11.61v.
The -12.00v is pretty far out of spec at -13.2-13.3v but the -5v seems pretty steady at -5.04.

Any help/advice will be appreciated. I'll be leaving town tomorrwow night (Saturday 21st) and won't be back till next weekend, but would appreciate any comments before then so it won't be driving me nuts while I'm gone. Again, thanks.

Do you have any other AMD CPU that is compatible with your NF7 that you could put in?

Hoot

I'm getting ready to hit the hay. If you don't have another CPU, unmount the 2600, inspect the pins for one the is bent or has debris on it. Inspect the socket pin holes as best you can. Inspect the core to make sure you didn't crack it due to uneven pressure when you were mounting the heatsink. If it looks okay, remount the 2600 and try again at default settings.

Hoot

I agree with Hoot.

Also, if you have another power supply, try that, as 11.55v on the 12+ line seems rather low. That's the line that the NF7-S draws the majority of power from for the CPU.

Could it be something like the graphics card overheating? that could also cause instability.

I have an Athlon 750, and a (defective?) 1700+ Palamino. I say it's defective because I could never get above 120-125FSB with it. That was on two different motherboards (the NF7-S and my old Epox 8KTA3) and also with two different psu's...the no name 300W that came with the case, and also the Arrow 500W. It was also attempted with 2 different sets of memory. 512MB of PC133 on the 8KTA3, and also on the NF7-S, along with the PC2700.

The more I think about it, the more I think that I possibly have had two psu's that just aren't up to it.

The only other psu I have is the old 300W that has no 4 pin 12V line. Will the NF7-S run without this?

yes it will. have a try.

If you have that 4 pin 12v line, it is better to use it, but if you don't, it will still work.

Ok, I'll give it a try tomorrw afternoon. I still lean towards the idea that it's the psu. I've mounted many, many Athlon/AthlonXp's, so I don't believe I introduced any debris or cracked the core, but if the other psu (which I tend to think is also weak) doesn't work, I'll have a look. Thanks for the advice.

Have I isolated the problem?

I got to thinking and you know I've already had the 1700+ Palamino in this mobo while I was waiting for my Barton. I'd posted a thread about how it couldn't run at 133FSB here:


http://forum.oc-forums.com/vb/showthread.php?s=&threadid=246412&highlight=1700+Palamino

The Palamino could never make it to Windows and be stable at anything over 120-125FSB. I got to thinking and underclocked the Barton. I took it down to 133FSB and it still wasn't stable. So I took it down to 100FSB and it was prime95 stable. I started boosting the FSB up and testing. It's Prime95 stable up to 120FSB.

In the other thread I also mentioned that the Palamino had the same problem on my old Epox mobo, and also on another NForce2 mobo besides this one. Before I got the NF7-S I had an FIC AU13, and it was the same deal. No go past 120-125FSB. Seems to me like it has to be the PSU. That's the only common denominator. Am I off base here, or does it make sense?

Two different chips, on two different mobos (one chip, the Palamino, on 3 different mobos), both CPU's with 2 different sets of memory (the Palamino with three different sets). Same problem in all configurations, the only thing all configurations have in common is the PSU.

Well firstly dont worry about the negative rails they are invariably inaccurate and are only for data comms and do not affect stability. That said your 12V rail is a major problem - and is more than likely the cause of your woes.

Yep, I think you're probably right. I have a friend who has a KT266A system with a Palamino 1600+. I think I'm gonna drop the Palamino off by his house and let him throw it into his system. If it runs up to speed in there, then that'll show 99.99% sure, I think, that the problem is the power supply.